Making sense of MaxEB

TL;DR - Ethereum’s core developer teams have agreed to include a code change in the next network upgrade (Pectra) that will increase the maximum amount of ETH from which a single validator can earn staking rewards (known as the maximum effective balance, or MaxEB) from 32 ETH to 2,048 ETH. Increasing MaxEB is arguably the biggest code change since Shapella initiated validator withdrawals last year, and is expected to encourage large-scale staking providers to consolidate their validators. This code change won’t affect the minimum amount of ETH needed to operate a validator, which will remain 32 ETH.

In the 18 months since Ethereum switched to a proof-of-stake consensus mechanism, the number of validators that have staked ETH to secure the network has grown to nearly one million.

That many validators should be a cause for celebration, right? After all, a massive validator set is a strong sign of decentralization and certainly provides robust economic security.

However, there are downsides to having that many computers talking to each other on a network. As the number of validators increases, so does the amount of messages that are being passed back and forth, causing congestion that slows down the entire network. It also causes unnecessary computing strain as each node is tasked with keeping their own running record of all network activity. Most believe those downsides outweigh any benefits of such a large validator set.

To address the situation, Ethereum’s core developers recently agreed to make a major code change to the consensus layer that is designed to significantly reduce the number of validators.

This update, proposed in EIP-7251, would increase the amount of ETH from which a single validator can earn rewards — known as a validator’s maximum effective balance, or MaxEB — from 32 ETH to 2,048 ETH. It’s designed to incentivize large-scale stakers and staking providers (like Coinbase) to consolidate their validators, which will reduce strain on the network while maintaining the same level of economic security.

The change is expected to be included in the next network upgrade, Pectra, scheduled to occur sometime before the end of the year.

As with any major change — and increasing MaxEB is arguably the biggest change since withdrawals were implemented in the Shapella upgrade last year — there are other considerations to keep in mind, a major one being how to address the existing initial slashing penalty, which if left untouched would expose validators with large effective balances to severe penalties that would likely discourage consolidation.

Let’s dig in.

Before we explore attempts to reduce the size of Ethereum’s validator set, we need to understand why the number of validators has steadily increased since the transition to proof-of-stake.

There are several contributing factors — including the proven stability of the network causing risk tolerance to grow among stakers, the rise of liquid staking, and innovations like Coinbase Wallet’s integration of partial ETH staking.

But, at the end of the day, the increasing number of people who want to stake their ETH are looking for one thing — yield.

Source: https://www.validatorqueue.com/

As an incentive for securing the network, validators earn staking rewards, which many think of as an annual percentage yield (APY) on their staked ETH. The amount of newly minted ETH validators receive (known as the issuance rate) is calculated as a function of the total amount of staked ETH — so, as more ETH is staked, the yield decreases. A year ago, in April 2023, validators had staked about 18 million ETH and were earning an APY around 4%. Today, there’s more than 31 million ETH staked, resulting in validators earning between 2.65% and 2.8%.

Despite this decrease in yield, staking still offers enough of a return — boosted by external sources of yield, such as airdrop farming and maximal extractable value (MEV) rewards — that it continues to attract validators.

As of April 23, there are more than 17,000 validators currently waiting in the queue to join the network, and only one waiting to exit. (Keep in mind, though, that almost twice as many validators are allowed to exit the network per epoch as can join, due to a decrease in the entry churn limit implemented as part of the recent Dencun upgrade.)

Source: https://www.validatorqueue.com/

Now that we've covered the basics behind the increased demand to become a validator, let's explore why all of that staked ETH needs to be spread across so many validators.

A validator’s maximum effective balance, or MaxEB, plays a key role in Ethereum's booming validator ecosystem. As the name implies, MaxEB establishes an upper limit for how much ETH each validator can effectively stake to earn rewards. 

The current MaxEB is 32 ETH. That means each validator can earn rewards (i.e., yield) from a maximum of 32 ETH — no more, no less (because a minimum of 32 ETH is still required to run a validator). 

This ceiling has unintended side effects that have contributed to validator bloat. A MaxEB of 32 ETH means any staker with more than that amount must operate multiple validators to earn yield on their full balance. (Remember, the 32 ETH minimum would require at least 64 ETH to spin up a second validator, 96 ETH for a third, and so on.) 

When applied to large-scale staking operators, which are potentially looking to stake thousands of ETH, it’s clear how this relatively low MaxEB has artificially inflated the number of validators on the network.

By increasing the MaxEB to 2,048 ETH, large-scale staking operators could earn yield from the same amount of staked ETH, but from a much smaller pool of validators.

Increasing a validator’s MaxEB to 2,048 ETH may sound like a simple way to incentivize consolidation, but there are other factors to consider, which could actually disincentivize consolidation if not carefully considered. 

Higher effective balance means increased slashing risk

To keep validators honest, their staked ETH is subject to being irrevocably taken away as punishment for not properly doing their job of validating transactions — this is known as “slashing.” (As a note, Coinbase’s Ethereum validators have never been slashed.)

There are a variety of slashing penalties encoded in Ethereum, including an initial penalty applied upon a validator’s first infraction. This initial slashing penalty is currently proportional to the validator’s effective balance.

Since 32 ETH is currently both the minimum amount needed to run a validator and the MaxEB, all validators face the same initial slashing penalty of 1 ETH, or 1/32 of its total stake. While the initial slashing penalty has never resulted in more than 1 ETH being slashed, the existing code has it scaling linearly, which means the larger a validator’s effective balance, the higher the penalty.

Can you see where this is going?

Scaled linearly, a consolidated validator with 2,048 ETH would face an initial slashing penalty of 64 ETH. 

The risk of losing 64 ETH, worth roughly $212,000 at the time of publication, would be tough for any validator to swallow. If left untouched, this risk would disincentivize consolidation among validators, undermining the original objective of increasing MaxEB.

How to address the initial slashing penalty is still under discussion among core developer teams. One proposal is to eliminate the penalty altogether, relying instead on the correlation penalty to deter bad behavior among validators. But the solution most likely to be adopted is to keep the penalty, but scale it back. Instead of a validator being slashed 1/32 of total stake, the new specs would result in an initial slashing penalty of 1/4,096 of total stake, or 0.5 ETH for a fully consolidated validator with 2,048 ETH staked.

For a deeper analysis of the initial slashing penalty vis-a-vis EIP-7251, read this research report from the Ethereum Foundation’s Mike Neuder.

Clogging up the exit and entry queues

If a staking operator needed to pull their staked ETH from one validator and apply it to a different validator, they would need to first remove the validator from the network, which requires joining the exit queue, before rejoining after a long wait in the entry queue. To note, only eight validators can join the network per epoch, a cap set via the recent Dencun network upgrade as a way to slow the growth of the validator set.

If large-scale staking operations began consolidating thousands of validators at once, the queues would become clogged very quickly, turning it into a cumbersome and costly process that could further disincentivize consolidation.

Because of this, part of the proposal would create a mechanism by which large-scale operators could consolidate “in-protocol,” meaning they could merge validators without needing to cycle through the exit and entry queues.

Reiterating a minimum effective balance

The MaxEB currently dictates that 32 ETH is both the minimum and maximum amount of ETH a validator needs to earn rewards.

To keep the minimum amount of ETH needed to run a validator at 32 ETH while increasing MaxEB, it’ll be necessary to create a new parameter in the consensus layer that fixes the minimum effective balance, which we think is safe to coin the MinEB. 

Impacts to solo stakers

So far we’ve only talked about how increasing MaxEB will impact large-scale staking operators, but the change also has important implications for solo stakers. 

A solo staker who owns 50 ETH and wants to earn yield on it by staking it to their own validator, simply cannot under current guidelines. The MaxEB means they’d only be able to earn yield on 32 ETH; and they couldn’t spin up a second validator, because they don’t have enough ETH.

Increasing the MaxEB would enable solo stakers to stake more flexible amounts of ETH. In our example, the solo staker would be able to stake all 50 ETH to their validator, earning yield on the entire amount.

As an added bonus, the yield earned from the initial 50 ETH will be able to remain staked on the validator, enabling the staker to earn compounding returns. (Currently, an automatic sweep takes any ETH above 32 on a validator and sends it to a wallet of the staker’s choosing.)

While increasing decentralization through the sheer size of the validator set sounds appealing, the network realities of surpassing one million nodes makes optimization necessary. The proposal to raise the MaxEB aims to streamline operations while preserving security.

However, simply increasing MaxEB without considering other implications (e.g., slashing penalties scaling linearly with higher balances, clogged queues, etc.) risks disincentivizing the validator consolidation the proposal seeks.

As with any major protocol update, balancing objectives requires evaluating all potential outcomes.

What’s Coinbase’s plan post-Pectra? As one of the large-scale staking operators this code change will impact, Coinbase is committed to doing what's best for the Ethereum network and community. At this time, and pending a final decision on how the initial slashing penalty will be handled, we believe that means consolidating our validators.

To go deeper on this topic, we’ll be hosting a webinar on April 25 with the Ethereum Foundation’s Mike Neuder, who co-authored the proposal to increase MaxEB.

Ben Rodriguez, Coinbase’s senior protocol specialist, will speak with Mike about the current state of affairs, why increasing MaxEB became the agreed-upon solution to reducing the size of the validator set, the implications of doing so, and whether there’s anything else to look out for in the upcoming Pectra update.